You are absolutely right. fsync was a problem while compiling. I fixed
that by passing -Dfsync=_commit
See for example:
On Mon, Jun 6, 2011 at 3:26 AM, Michael Schumacher <schum...@gmx.de> wrote:
>> > On Sat, Jun 4, 2011 at 12:14 PM, Partha Bagchi <parth...@gmail.com>
>> >> I tried again. The error still persists:
>> >> gegl-buffer-save.c: In function 'gegl_buffer_save':
>> >> gegl-buffer-save.c:227: error: 'S_IRGRP' undeclared (first use in
>> Von: Partha Bagchi <parth...@gmail.com>
>> OK. I create a patch using diff. Would that be sufficient? I attached
>> it to this mail. Hope it comes through.
> The S_*GRP and S_*OTH flags aren't defined on the Windows platform, and to my
> knowledge the bits associated to them don't have any meaning there anyway.
> I've committed a change to GEGL to fix this (a #ifdef check to change the
> open command to not use them on the Windows platform).
> The next problem that you'll most likely encounter is the undefined fsync in
> gegl/buffer - in my local tree, I've added a AC_CHECK_FUNS(fsync) directive
> to my configure.ac and wrapped the fsync calls in #if HAVE_FSYNC.
> I guess that fsync could be redefined to FlushFileBuffers (see
> http://www.humboldt.co.uk/2009/03/fsync-across-platforms.html) on the Windows
> platform unless a fsync implementation is found by the check.
> Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
> belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de
> Gimp-developer mailing list
Gimp-developer mailing list